I think what GP is saying is that even with large, multi-million margins in the popular vote, the result can still be "close" because many of those votes count for nothing in the electoral college. Hillary Clinton could have received four million fewer votes in California 2016 and she would still have won that state while losing the electoral college. But if she had merely won an extra ~100,00 votes in a few crucial swing states like Michigan and Pennsylvania, America might have had its first female president.

I haven't checked the numbers but I think GP is arguing that the Biden-Trump result was "closer" in this sense than the Clinton-Trump result was; i.e. the margin in the electoral college and nationwide popular vote may have been wider, but 2020 was still a closer call in the small number of swing states that might have actually changed the overall result.